Population Geneticist/Conservation Biologist Chicago Zoological Society

1(5)

Comments on the genetic issues related to the new Action Plan for the Lesser White fronted Goose (L WfG)

Robert C Lacy, PhD

Population Geneticist/Conservation Biologist Chicago Zoological Society

Chair

IUCN/SSC Conservation Breeding Specialist Group

Committee on Evolutionary BiologyUniversity of Chicago

I will preface my remarks by stating that I have not before been involved in any of the discussions or analyses of the LWfG or any related species. My comments below are in response to the set of documents sent to me by Sergey Dereliev of the UNEP/AEWA Secretariat.

My background is that I was trained as an evolutionary biologist, with work in population genetics, ecological genetics, and behavioral genetics. I have worked for the past 20 years as a conservation geneticist for the Chicago Zoological Society, with adjunct faculty positions at the University of Chicago and University of Illinois. My research has included: experimental studies of the effects of inbreeding and intercrossing on Peromyscus mice; analyses of the genetic changes and inbreeding effects that occur in captive breeding programs for wildlife species; development of statistical techniques for pedigree analysis and the management of breeding programs; and development of computer simulation models for population viability analysis for assessing threats to wildlife populations and testing the likely impacts of proposed management actions. I have taught short courses to wildlife managers and zoo biologists on the genetic management of endangered species. For the past 3 years, l have served as the chairman of the IUCN/SSC Conservation Breeding Specialist Group - the network of experts who provide technical assistance on matters related to use of captive breeding programs to serve species conservation, and related programs of intensive population management. I have provided advice to government agencies on the genetic management and recovery plans for whooping cranes, Puerto Rican parrots, and three penguin species, and also for many species of mammals (e.g., black-footed ferrets, beach mice, and Florida panthers in the USA, eastern barred bandicoots and Leadbeater's possums in Australia, and all five extant species of rhinoceros), and a few reptile and amphibian species.

It is not clear to me if the primary disagreement about the genetic issues related to conservation actions for the LWfG are due to different opinions about the genetic data and analyses, or to different interpretations of the implications of those data for conservation, or to both the data and the conservation implications. With respect to the data themselves, it seems to me that with the most recent molecular genetic analyses, the genetic characterization of the LWfG is becoming clear (although I expect that some of those involved in the debates may still disagree with parts of my description of the information now available).

The mitochondrial DNA data show that two divergent clusters (each with a primary common type and a number of variants that differ only by one or two mutations of likely regent origin) of mtDNA haplotypes occur in the wild populations of LWfG, and an additional two general types occur in the birds in the captive breeding programs for the LWfG. The two general types (West and East) found in the wild LWfG both exist in all wild populations, but at different frequencies, although some sub-types (slight variants that would represent recent evolutionary changes) of the W and E types are unique to one region or the other. The other two general forms of mtDNA observed in the captive geese have been found to be typical of the Greater White-fronted Goose (GWfG) and the Greylag Goose. The sampling of LWfG from wild populations has been sufficiently extensive so that it is very unlikely that both the typical (E and W) LWfG and the typical GWfG forms of mtDNA are prevalent in the natural populations of LWfG (as could have occurred if both forms persisted in the LWfG from an ancestral population that preceded the evolutionary split between the LWfG and the GWfG). In addition, although the numbers of LWfG in the wild populations has been in decline, the numbers are not so low that it would have been possible that once common mtDNA haplotypes would have been lost from the wild populations but still persisted in non-hybridized captive flocks. Even if the wild populations had lost some mtDNA haplotypes that persisted in captive flocks, it is not plausible that all the types characteristic of the GWfG (and the Greylag Goose) would have been lost - loss of haplotypes from small wild populations would be expected to have been more random. Thus, the mtDNA data do show that the captive stocks of LWfG have been hybridized with two other species.

Mitochondrial DNA are inherited only from the maternal parent, so the data on mtDNA haplotypes can show that hybridization occurred, but not how muck occurred. Birds labeled as LWfG would show mtDNA haplotypes characteristic of other species only if their maternal lineage (mother, grand-mother, etc.) descended from the other species. Breeding between a male GWfG (or a hybrid) and a female LWfG would not be detectable by this method. Variants of nuclear genes can be used to detect ancestry through the paternal side, and can be used to quantify the average amount of genetic ancestry in a hybrid population that descends from each source species. The RAPD technique can reveal species-typical DNA patterns. However, the technique relies on non-specific DNA probes (i.e., sequences of DNA that bind, with uncertain fidelity, to unknown numbers of genes in each species), so that the repeatability and interpretation of those data are orten uncertain. For these reasons, most geneticists are willing to use RAPD data to suggest possible patterns, but are unwilling to use them to provide rigorous quantitative estimates of population parameters - such as the degree of divergence between two populations or extent of hybridization in a possibly mixed population.

Microsatellite DNA markers (sections of repeated short sequences of DNA) provide more repeatable and precise estimates of population differences, because - if proper precautions are taken - we can confirm that the variants at each stored locus are simple alleles that follow Mendelian inheritance. The recent work by Ruokonen et al. assessed 10 microsatellite loci - sufficient to document that a number of captive LWfG (including some that had a mtDNA haplotype typical of LWfG) contain evidence of GWfG ancestry. Considering both types of genetic evidence, at least 36% of the captive LWfG that were analyzed were shown to have some hybrid ancestry. The close evolutionary relationship and consequent overlap of nuclear genetic alleles prevented the researchers from quantifying the proportion of GWfG ancestry in the captive stocks, but the above numbers support the view of Ruokonen et al that the present

captive stocks are "unsuitable for further reintroductions or supplementation." Rigorous testing of the mtDNA and microsatellite DNA of captive birds (with, preferably, an increase in the number of microsatellite loci scored) could allow selection of birds in the captive stocks that have low probability of hybrid ancestry, but without at least 3-4 diagnostic nuclear loci (none are yet known) or good pedigree records (apparently not available for the captive stocks), it would not be possible to select a subset of captive birds that exclude all hybrid ancestry.

The combination of mtDNA and nuclear DNA data are now showing a clear pattern of moderate but not strong genetic divergence among wild populations of LWfG. The lack of sharp discontinuities in the allele frequencies and the estimated numbers of migrants that would result in the observed differences in allele frequencies indicate that there is (or recently has been) enough movement of LWfG between eastern, central, and western parts of the species range to have prevented evolutionary divergence and also to have prevented extreme loss of genetic diversity and accumulated inbreeding within any population segment. Thus, the populations do not appear to be genetically isolated to the extent that they would be considered to be evolutionarily significant units or subspecies. The populations may have diverged partially with respect to traits adapted to local conditions, but the genetic mixing makes it unlikely that important adaptive differences have become "fixed" in (i.e., unique to) segments of the species range. Thus, dispersing or translocated individuals may have lower fitness because they may more often have genotypes best suited for a different habitat, but each population probably still contains the range of genetic variability necessary to adapt to local conditions.

The populations in Fennoscandia appear to have some reduction in genetic variation relative to more eastern populations, but there is not yet evidence of problems arising from inbreeding, andsuch problems would not be likely to accumulate rapidly, given the evidence for some genetic connections to the larger populations to the east. Thus, it does not seem to me that it is necessary at this time to release individuals in Fennoscandia in order to "rescue" the population from a lack of genetic diversity.

Although I do not think that the evidence suggests a current need to provide genetic rescue of the Fennoscandian population of LWfG, I do not agree with the suggestion that restoration of genetic variation should wait until the Fennoscandian population is extinct. Release of birds from other sources (whether from captive flocks of documented origin or translocations from other wild populations) may shift allele frequencies, but given the genetic closeness of the LWfG populations in different regions it is hard to see how such releases could disrupt local adaptations to the extent that it would damage the prospects for the population. Instead, the effects of such releases would be to restore genetic variants that could have been lost from the small population and to reverse local inbreeding. Moreover, the extent of disruption of any local adaptations would be greatest if the remnant population is allowed to become nearly extinct before genetic management was resumed. Waiting until the local population is extinct would actually ensure that any local adaptations that did exist would be lost, instead of remaining within a more variable gene pool that could continue to adapt to local conditions.

In contrast to the lack of evidence of notable genetic isolation of the Fennoscandia population, the extent of divergence of frequencies of genetic alleles does indicate that inter-populational dispersal is rare enough that the populations are demographically independent (or nearly so) and

should be considered to be separate conservation "management units." Thus, the movement of individuals into the Fennoscandia population is not sufficient to provide significant demographic reinforcement of a declining population nor reestablishment of a population following regional extirpation. This is especially so if, as suggested from the mtDNA patterns, most dispersal between regions is by males, with females being more philopatric. Dispersing males are as useful as are females for preventing genetic isolation and inbreeding, but they have little demographic impact. The fact that the population in Fennoscandia continues to decline is evidence that natural dispersal among regions is not sufficient to support that population if it is not protected as an independently vulnerable management unit.

There is a difference of opinion among the experts regarding whether the small and declining wild population in Fennoscandia is doomed to extinction if it is not supplemented. I have been involved with developing and assessing population viability models for a number of endangered species (but not for the LWfG). The probability of population recovery - after the causes of decline are removed - is a function of the population size, with very small populations being more likely to experience inbreeding depression, locally imbalanced sex ratios and other difficulties in finding mates, vulnerability to disease epidemics or other local catastrophes, and other problems intrinsic to small populations. The size of population below which extinction becomes likely varies among species, based on life history, habitat characteristics, evolutionary history, and other factors. It is perhaps misleading to consider any given number to be a "critical" population size, as smaller populations are at greater risk, but there is no size below which a certainty of persistence changes to a certainty of extinction. However, for any given species and environment, the relationship between population size and extinction probability is amenable to analysis.

For relatively long-lived vertebrates (such as geese and most birds), I do not believe that the numbers that currently exist in the wild population of LWfG in Fennoscandia would allow classification of the population as either "doomed" or "safe". (I.e., both sides of the debate seem to have overstated their case.) Many populations have recovered from even lower numbers, such as the whooping crane recovering steadily from a low of only N= 15, after protective measures were implemented. However, the whooping cranes did suffer a significant loss of genetic diversity, and this is likely a cause of the observed high rate of genetic anomalies of development and high susceptibility to some diseases. If the current population of about 20-30 breeding pairs of LWfG is so low as to make damaging genetic impoverishment inevitable, then almost all captive populations of wildlife species would have to be considered to have no conservation value, as rarely are the captive stocks founded with more than 25-30 breeders. Fortunately, not very much genetic diversity is lost when a population goes through a bottleneck of about 20 pairs for one or a few generations. For example, 25 randomly breeding pairs would lose about 1% of its gene diversity (heterozygosity) per generation, allowing it to persist for 10 generations before it lost the 10% of gene diversity that has often been considered to be level of concern for stocks of wildlife or domesticated species. (Often, however, some pairs are muck more productive than others, rather than there being a random distribution of breeding success, so actual losses of genetic diversity might be about twice this rate.)

On the other hand, we should not have confidence that the population of LWfG in Fennoscandia can recover without assistance. First, the current steady decline must be stopped, or else all other

conservation actions will provide at best only temporary assistance. After stopping the decline due apparently to hunting mortality, the existing population may or may not be able to recover without supplementation. The persistence to today of apparently a single remnant male ivorybilled woodpecker and other examples of presumed species losses that have been avoided (or delayed) should not be taken to be evidence that that species or any species can recover from very low numbers. Florida panthers declined to perhaps only 10-20 breeding individuals for several generations, and the severe inbreeding effects were reversed only after intercrossing with another population. Black-footed ferrets had been presumed to have been rescued after a decline to only about 10 unrelated animals (and their offspring), but they are now showing declining reproductive success that most likely results from the inbreeding that occurred in the population bottleneck. The wild population of LWfG is approaching the level at which we might soon see dangerous effects of inbreeding, but the population should still be recoverable, especially if occasional natural or manipulated immigration from central and eastern populations occurs.

If a captive stock is used for supplementation of the wild LWfG, it would be wise (in light of the data discussed above) to initiate that stock with birds that are "pure" LWfG. Starting new stocks from birds captured in Fennoscandia or more eastern populations might be costly, but perhaps no more so than the extensive genetic testing that would be needed to derive a pure or largely pure population from existing captive stocks. In addition, existing captive stocks have not been managed to minimize genetic changes, so they may have adapted genetically to captivity in ways that include loss of species-typical breeding preferences that serve as isolating mechanisms. After a population is established, monitoring and genetic management of a captive population is not much more difficult or costly than maintaining a population without attention to the pedigree, and can increase the genetic effectiveness of a breeding population several-fold relative to a stock that is not managed genetically. (I.e., a stock managed with the methods used for wildlife species in well managed breeding programs can lose genetic diversity as slowly as would an unmanaged population that is two or three times larger.)

Perhaps the most difficult issue facing the conservation and management authorities is to decide what to do with already released birds (and their descendants) that carry non-LWfG genes. It may not be possible to remove these birds or the hybridized genomes from the wild, especially if they have already further interbred with the remnant wild population. It is possible that species-isolating mechanisms have broken down in the hybrids, so that the released birds and their descendants might now provide a path for continued introgression of genes from GWfG into LWfG populations. Otherwise, the extent of introgression of non-LWfG genes into Fennoscandian populations is probably not so great that it will do long-term damage to the ecological and evolutionary future of LWfG in Fennoscandia. Very small amounts of gene flow from closely related species is not an uncommon occurrence in natural populations. Future releases of documented LWfG, occasional immigration from central and eastern populations, and natural selection could all serve to slowly reduce the level of genetic contamination of the LWfG and restore the species to a genetically more natural condition.